[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian/sparc problems (and solutions?)



I elaborated on this earlier, but I decided to be very stupid this morning and
didn't bother cc-ing the mailing list.  Oops.  So, once more:

> Well.. in 2.0.35, on my SS2 (and I know others have experienced this) you
> cannot suspend processes with ^Z in bash.  More accurately, you can suspend
> the process, but bash never comes to the foreground - so you are left with a
> dead terminal sitting on a suspended process.

This is also a problem in ksh, so it's probably not just a bash thing.  Since
I upgraded to the latest telnetd, though, this has been less of a concern.  If
I log in from console or telnet in ^Z works fine (and returns a shell prompt),
but if I ssh into the machine ^Z still isn't coming to the foreground after
stopping the process.  Perhaps ssh should be recompiled?

> I don't personally use modprobe - those problems were Michael's, perhaps
> he'll elaborate.

I do use modprobe.  Until I rebuilt modutils I was getting an "unknown error"
loading modules, with a cryptic (probably a memory addr) error number.  The
modutils currently on master fix this.  Things are working fine now.

> > 2.0.95 broke libstdc++2.8 and there's no source to recompile (see
> > debian-devel thread).

libstdc++2.9 seems to work fine.  Why not just recompile all the packages
linked against 2.8?

> >> And then we have the big kernel version problem.  Any recent 2.1 kernel
> >> (later than eightysomething ????) will not run our bash.  Or not last I
> >> tried, anyhow.

2.1.x kernels don't run ksh either.  Perhaps this is some conflict with libc?

> ldd doesn't work on shared libs, which breaks dpkg-shlibdeps.  This is
> apparently SS2 specific, and fixed in kernel 2.1.x (x>80 or so).

I'm using an ss4 and I don't see it, so it probably is ss2 related.

> Recently, update-menus has started segfaulting on me.  I have recompiled the
> menu package against stdc++ 2.9, but the problem remains.

I built menu a while ago & linked it against libstdc++2.9.  I haven't been
seeing any unusual segfaults.  Perhaps this is an ss2 problem also?

Just a random thought - lots of things have been experiencing quirky behavior
since glibc was upgraded.  Recompiling for the latest snapshot usually helps
a lot.  X was built a while ago & glibc has been upgraded since then.  Maybe
someone should recompile it and see if it runs smoother.

-- 
Mike "No .sig for you" Shuey


Reply to: